First Steps Into the Project

- Reading the tabular data
- Reading the raster file and the information it contains
- Displaying the tiles on the raster file 

Setting working directory

Modules and Libraries Needed

Tabular Data

Some information (especially on elevation and slope of the land) appears on the tabular data but is not on the raster file 
Let's visualize that data using the shapefile 

Shapefile

We merge the data of interest to the shapefile from the analytical table using the primary key [tile_h, tile_v]

Raster File

Properties of the Raster File

2-Band Raster

In displaying the two bands, we can see that Band1 gives us the landcover code while Band2 is just an indicator if the pixel is on the state or outside of it

We can also see that the origin of the file is at the top left corner

Let's take a closer look at Band1

Band1 contains the data we need since the tiles always encompass at least some "instate pixels", let us join it with "landcover_map" to look at the spatial distribution of land types across the state

How do the tiles cover the state ?

This is a supervised learning problem and our observations are the tiles covering the state. Let us plot these tiles on top of the raster file for visualisation purposes. For each blue square that doesn't have a target we will have to predict how many pixels will belong to one of these 4 categories in 2016:
    - developed_open_space
    - developed_low_intensity
    - developed_medium_intensity
    - developed_high_intensity